Side by side comparison with CUDA and OpenCL
epsi1on opened this issue · comments
Hi,
Is your feature request related to a problem? Please describe.
Is there any side by side comparison with well known solutions like CUDA available for this library? like for well known parallel problems (matrix add or matrix multiplication).
i would like to know how the ILGPU's performance is compared to a native cuda or OpenCL code.
Thanks for great work
Describe the solution you'd like
A page on wiki, which compares the runtime speed of same algorithm on same input data, in cuda, opencl and ilgpu.
In order to user can compare the ILGPU with cuda, and help select ilgpu if the speed is reasonable compared to cuda.